The $145K AI Security Engineer Role. No Degree. No Cert.

The Problem

Most AI security education will not get you hired.

AI Security Engineer is the fastest-growing role in cybersecurity. Every company with an AI product needs one. The job listings are out there. The salaries are real. The framework hiring managers screen on is published (OWASP LLM Top 10, MITRE ATLAS, NIST AI RMF).

But the people trying to break in are stuck on the same wall: no portfolio. No build. No proof. Just a stack of slides and quiz scores from courses that never asked them to write a line of code, deploy a system, or attack their own work.

Lecture-heavy theory courses. Hours of slides on prompt injection. Zero hours building an actual AI system to defend.
The single-pillar trap. Most courses teach LLM security and stop there. Real AI systems have four attack surfaces. Cover one and you cover nothing.
No production reality. No cloud foundation. No API middleware. No SOC observability. The pieces that take an AI build from demo to defendable get skipped.
No portfolio artifact. You finish the course and have nothing to show a hiring manager. Just a PDF certificate they will not open.

How to break into AI Security as an engineer

Why Now

AI Security Engineer is the fastest-growing role in cybersecurity.

Every company deploying an AI product needs someone who can defend it. The job titles are real. The salaries are real. The framework is published. The opportunity is now, before the field saturates.

Why now is the moment to enter AI Security

The Solution

AI Security Engineer Light is the build-first answer.

Four production AI systems you build, attack, and harden yourself. Every project follows the same loop: build it, release it, break it, harden it. Working code. Real attacks against your own builds. Before-and-after evidence files. The portfolio artifact you push to GitHub and link from your resume.

You build 4 real AI systems across the 4 pillars: LLM Security, AI APIs and MCP, Agentic AI, and Vibe Coding.
You attack each one yourself using documented attack scripts. Prompt injection, training data poisoning, tool misuse, memory poisoning, cost runaway, signature forging.
You harden each one with NeMo Guardrails, OAuth audience binding, signed memory, JSON Schema validation, and 5 Sigma detection rules mapped to MITRE ATLAS.
You deploy the cloud and API foundation every pillar sits on. AWS Terraform baseline. FastAPI middleware. Wazuh and OpenSearch.
You push the build to GitHub as your portfolio artifact. With the threat model, the attack evidence, and the harden walkthrough alongside it.
You learn the job-search playbook. The 30/60/90 day plan. Outreach scripts. How to talk about your build in interviews.

The Curriculum

The 4 Pillars of AI Security.

AI security is not one domain. It is four overlapping problem spaces, each with its own threat model. Most teams cover one or two and miss the others. This is the framework that lets you defend any AI system.

Pillar 1 · What You Build

LLM Security

RAG bot with NeMo Guardrails. 4 attack scripts: prompt injection, data leakage, training data poisoning, hallucination chains. Full harden walkthrough.

Pillar 2 · What You Build

AI APIs and MCP

Bedrock API + MCP server hardened against 9 STRIDE attacks. Per-token cost caps. Audience-bound OAuth. The deployment layer most teams get wrong.

Pillar 3 · What You Build

Agentic AI

LangGraph multi-agent build with signed memory and identity propagation. Defends against tool misuse, memory poisoning, cascading hallucinations, privilege compromise.

Pillar 4 · What You Build

Vibe Coding

The defensive pipeline for AI-assisted development. Catches insecure code generation, hardcoded secrets, dangerous functions, slopsquatting attacks.

The On-Ramp

The first 5 sessions get you to your first pillar build pushed to GitHub.

The honest timeline to interview-ready across all 4 pillars is 6 to 8 weeks with cybersecurity or cloud experience, 12 to 16 weeks from zero. Inside that timeline, the on-ramp is fast. About 8 hours of focused build across these 5 sessions and you have a working pillar build, attacked and hardened, pushed to GitHub.

Session 1 · 90 min

Cloud foundation

Chapter 5. Terraform up an AWS sandbox. Tear it down. Cost: $0 between sessions.

Session 2 · 2 hr

API + SOC layer

Chapters 6 and 7. Wazuh, OpenSearch, FastAPI middleware patterns.

Session 3 · 3 hr

Pick a pillar & build

Most start with Pillar 1 (LLM Security with NeMo Guardrails). Run all 4 attack scripts.

Session 4 · 90 min

Portfolio artifact

Chapter 13. Push the build + attacks + threat model to GitHub.

Session 5 · 60 min

30/60/90 day plan

Chapter 16. The outreach scripts and timeline to interview-ready.

After the on-ramp, the next 5 to 7 weeks deepen each pillar with more attack scripts, harden walkthroughs, and the interview-prep material in Chapters 14 to 16. The work compounds. Every session is more portfolio.

Everything Inside

What you walk away with.

Chapter 1. The AI Security Engineer role: what you actually do, who pays you to do it, the framework hiring managers screen on.
Chapter 2. The 4 Pillars threat model: how to map any AI system to its attack surfaces.
Chapter 3. OWASP LLM Top 10, MITRE ATLAS, and NIST AI RMF: the references you cite in production and in interviews.
Chapter 4. The build-it-release-it-break-it-harden-it methodology and why courses without it produce candidates who cannot get hired.
Chapter 5. Cloud Security Foundations: AWS account baseline with Terraform. Deployable. Cost-honest. Includes the teardown workflow.
Chapter 6. API Security Foundations: OAuth audience binding, JSON Schema, per-token cost caps, audit logging. Working FastAPI middleware code.
Chapter 7. Observability and Detection: Wazuh + OpenSearch deployment, 5 Sigma rules with MITRE ATLAS mappings. The SOC side most courses skip.
Chapter 8. Pillar 1 LLM Security: NeMo Guardrails RAG bot, 4 attack scripts, harden walkthrough.
Chapter 9. Pillar 2 AI APIs and MCP: Bedrock API + MCP server hardened against 9 STRIDE attacks.
Chapter 10. Pillar 3 Agentic AI: LangGraph multi-agent with signed memory and identity propagation.
Chapter 11. Pillar 4 Vibe Coding: defensive pipeline against AI-generated insecure code and slopsquatting.
Chapter 12. Communicating risk: how to write a one-page threat model a CISO will actually read.
Chapter 13. The Portfolio Artifact: how to package your build for GitHub so a hiring manager opens it.
Chapter 14. Interview prep: the 12 technical questions every AI security interview includes, with model answers from your own build.
Chapter 15. Outreach playbook: how to get the interviews. Specific scripts. Specific channels.
Chapter 16. The 30/60/90 day plan: from purchase to first AI Security Engineer interview.

Who Is Teaching This

The work behind the teaching.

Zach Marcy, Cybersecurity Architect and Mentor

I am Zach Marcy, online HackWithZach. I am a Cybersecurity Architect and Mentor with 20+ years of IT experience and 6 in cybersecurity. My day-to-day work is designing and securing cloud environments that deploy and secure APIs and AI.

HackWithZach is the project where I turn that work into education for people breaking into and growing within cybersecurity. SOC operations, SOC and cybersecurity architecture. Quality over quantity, depth over noise, real-world relevance above all else.

Every project in this course is something I have built, attacked, or hardened in production. Every Sigma rule comes from a real detection use case. Every attack script reflects how AI systems actually get compromised. Not theory. Not slideware. Production reality, written down so you can build the same thing.

Cybersecurity Education That Gets You Hired, Promoted and Paid.

Why This Is Different

AI Security Engineer Light vs. everything else.

There are AI security courses on every platform now. Here is what makes this one the one that gets you hired.

What you get	AI Security Engineer Light	Typical AI security course
Working code you build yourself	Yes. 4 full pillar projects.	Slides and quizzes.
Real attacks against your own builds	Yes. Documented attack scripts per pillar.	Read about attacks. Run none.
Cloud foundation included	Yes. Terraform-deployable AWS baseline.	Skipped.
API security middleware code	Yes. FastAPI patterns.	Skipped.
SOC observability layer	Yes. Wazuh + OpenSearch + Sigma rules.	Skipped.
Covers all 4 AI attack surfaces	Yes. LLM, AI APIs, Agentic, Vibe Coding.	Usually LLM only.
Portfolio artifact you push to GitHub	Yes. With the threat model and attack evidence.	PDF certificate.
Job-search playbook	Yes. Chapter 14, 15, 16.	Skipped.
Mapped to OWASP LLM Top 10, MITRE ATLAS, NIST AI RMF	Yes. Explicitly.	Sometimes mentioned.
Written by someone defending AI systems in production	Yes.	Often by someone teaching what they read.

After You Finish

What you can do that you could not do before.

Defend an AI system in production. Not in theory. You have built one, attacked it, and hardened it.
Read a job description for AI Security Engineer and know exactly which sections of your build match each requirement.
Walk into an interview with a public GitHub repo a hiring manager has already opened.
Write a one-page threat model a CISO will actually read.
Map any AI system to the 4 Pillars and surface the threats no one else on the team sees.
Speak fluently about OWASP LLM Top 10, MITRE ATLAS, and NIST AI RMF in technical conversations.
Position yourself for the AI Security Engineer role, the Cloud Security Engineer role with an AI specialization, or the SOC Analyst role at a company running AI in production.

The market reality.

This is what the role pays and where the work is. Source data from Glassdoor.

AI Security Engineer total compensation: $145K to $235K. Entry $90K-$130K. Senior $200K-$350K+. Source: Glassdoor.

The Founding Member Community

30 founders. One private Discord. Direct access to me.

The first 30 buyers of AI Security Engineer Light get more than the course. You get the Founding Member Discord: a private channel limited to the 30 of you, with direct access to me. Ask questions on your build. Share threat models for feedback. Get your portfolio reviewed before you push to GitHub. No rigid cohort schedule, no group calls to show up to, no instructor watching the clock. Just builders helping builders, available whenever you need it.

Direct access to Zach. The same architect who wrote the course is in the channel answering build questions.
Portfolio reviews before GitHub push. Drop your repo link, get specific feedback on what a hiring manager will see.
Threat-model feedback. Share the STRIDE table or the one-page model you wrote in Chapter 12, get it sharpened before you put it in front of an interviewer.
The other 29 founders. Working through the same builds, hitting the same walls, sharing what unstuck them. The accountability layer a solo PDF cannot give you.
Founding 30 only. When the 30 seats fill, the founding channel closes to new members forever. The 31st buyer enters the broader member Discord at $197.

Founding Member Release

AI Security Engineer Light

$27

Normal price: $197

First 30 founding members at $27. Price moves to $197 permanently when the 30 founding seats fill. This is the only time you can buy AI Security Engineer Light at $27.

The complete AI Security Engineer Light build blueprint, instantly downloadable
4 production AI systems you build, attack, and harden yourself
The cloud foundation, the API security middleware, the SOC observability layer that every pillar sits on
The portfolio artifact playbook a hiring manager opens before they interview you
The 30/60/90 day plan from purchase to first AI Security Engineer interview
Lifetime access and free updates as AI security evolves
Founding Member Discord: private channel limited to the first 30 buyers, direct access to Zach, portfolio reviews before you push to GitHub
Founding member status: $27 price locked, never offered again

Claim Your Founding Member Seat — $27

Secure checkout via Stripe. Instant download upon payment. All sales final on the $27 founding member tier. By completing your purchase you acknowledge the no-refund policy.

FAQ

Questions I get before people buy.

How long does this actually take?

Honest answer: 6 to 8 weeks of consistent work if you already have cybersecurity or cloud experience. 12 to 16 weeks if you are starting from zero with no prior security or cloud background. The 5-session on-ramp gets you to your first pillar pushed to GitHub in about 8 hours of focused build. After that, the next 5 to 7 weeks deepen each pillar to interview-grade. No shortcuts. No fake "transform your career in 7 days" promises. Real work, real timeline, real outcome.

Do I need an advanced degree or another certification?

No. The role does not require either, and this course does not require either. You will not be told to go grind for a vendor exam before you can do the work. The cloud and API foundations chapters give you everything you need to do the pillar builds. The portfolio artifact is what gets you in the interview, not a piece of paper.

Is the $27 price actually limited to 30 people?

Yes. Founding member pricing is a one-time release for the first 30 buyers. The published price is $197, and the page moves to $197 when the 30 founding seats fill. If you are reading this and the page still shows $27, the seats are still available.

Do I need a CS degree?

No. The course is built for working IT professionals, SOC analysts, and security engineers pivoting into AI security. The cloud and API foundations chapters give you everything you need to do the pillar builds.

Do I need an AWS account?

Yes, ideally a sandbox sub-account. The Terraform baseline costs roughly $11/month while running and goes to $0 when you tear it down between work sessions. Chapter 5 includes the honest cost breakdown and the teardown workflow.

Is this video?

This is the PDF version: text, code blocks, screenshots, and diagrams, designed for read-and-build. Video walkthroughs of every pillar build are part of the upcoming Full tier (target September 2026). If you want video, see the Full waitlist.

Refunds?

All sales are final on the $27 founding member tier. The founding pricing is 86% off the standard $197 price and is reserved for buyers ready to commit to the work. To evaluate before you buy, grab the Free Training (the 4 Pillars PDF) at hackwithzach.com/free-training and read this page in full so you know exactly what you are getting.

What if I want the videos and the monthly calls later?

The Full tier (target September 2026) adds recorded video walkthroughs, monthly live calls, full GitHub repos, and the broader Full-member Discord. The Upgrade tier credits your Light purchase against the Full price, so you only pay the difference. As a founding member, you also keep access to your Founding Member Discord. Get on the Full waitlist here.

What is the Founding Member Discord?

A private Discord channel limited to the first 30 buyers of AI Security Engineer Light. You get direct access to Zach for build questions, portfolio reviews before you push to GitHub, threat-model feedback, and the other 29 founders working through the same material on the same timeline. When the 30 founding seats fill, the channel closes to new members. Buyer 31 (paying $197) enters the broader Light-member Discord launching alongside Full tier in September 2026.

Land the $145K AI Security Engineer role. No degree. No cert.